# Shared

> Shared runtime and evidence references

- Skill: `hashgraph-online/shared` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add hashgraph-online/shared`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hashgraph-online/shared/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: hashgraph-online (https://skillmd.com/u/hashgraph-online)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/hashgraph-online/shared

---


# Shared References

Shared files describe runtime capabilities and evidence formats. They are
context, not permission to start a runtime, tracker, substrate, network call,
or external mutation.

Just-in-time loading works because a reference read only when a consuming
skill needs it cannot silently become a dependency; anything loaded by default
eventually gets treated as one.

Named failure mode — **reference promotion**: shared prose quietly outranking
a source skill contract because it was read more recently.

Anti-pattern: citing a shared file as authority for starting a tool or
runtime. Corrective: authority comes from the caller or the consuming skill's
contract; shared files only describe.

- Default to the current agent and local shell.
- Use a runtime-native fresh context only when the caller or consuming workflow
  requests it.
- Treat runtime and factory state as adapter evidence; never translate it into
  core Plan, Candidate, RPI, or verdict state.
- Missing optional tools degrade only the optional capability that needs them.
- Source skill contracts and executable behavior outrank shared prose.

The core loop has no hard dependency on this library.

---

**Source:** [`hashgraph-online/awesome-codex-plugins`](https://github.com/hashgraph-online/awesome-codex-plugins) → `plugins/boshu2/agentops/skills-codex/shared/SKILL.md`